This research investigates the effects of stimulus variables on contrast interactions during operant discrimination training in pigeons. These effects seem analogous to the induction effects observed by Pavlov during classical conditioning. Induction provided evidence for Pavlov's theory of discrimination learning based on conditioned excitation and inhibition. Many researchers feel that this model may be useful in accounting for the effects of discrimination training during operant, as well as during classical, conditioning. Traditional operant mechanisms do seem incapable of accounting for discrimination learning; thus, a change in our views of operant behavior does seem in order. However, the Pavlovian model is not the answer. This proposal summarizes evidence which indicates that this account, or any other which relies on either absolute or relative excitation and inhibition, cannot account for contrast or for discrimination learning in general. The research proposed, using maintained generalization gradients, should help determine the factors which underlie contrast; in so doing, it will contribute to a better understanding of operant behavior and perhaps, classically conditioned behavior as well.